Keeping rpm constant, obviously, intake boost won’t accomplish anything since fuel efficiency would be determined by the lack of cylinder filling, so what ever boost was attained would have to be negated by the throttle body position. The gain would have to come from the exhaust side, specifically, intake contamination.
As an example, keeping the cylinder filling from the intake now constant also and looking at the power produced, let’s say that the cylinder is filled 25%. The work needed is a function of compressing that 25%, the work generated, given a 4x Kelvin rise in heat after ignition would be four times that, we’ll call it a net work output of (4-1) =3. Now let’s say that the restriction on the exhaust side leaves some exhaust in the cylinder, filling another 25% of that cylinder, the work of the compression then doubles, but the work generated also increases, since both pressures have increased. The net result on the positive work side would be that the final temperature would be lower, while the pressure would be slightly higher:
The relative work then needed would equal 2, the relative work output would equal 5, giving the same amount of net work, but some of the heat would be absorbed by the remaining exhaust charge in lieu of the cylinder and head. So the net result, since there is a lower temperature differential between the gas and it's surroundings, would be that the heat would be used for expansion of the exhaust gas instead of heating the engine. Unlike an egr system, which also increases mileage, the restriction of exhaust gasses would also suffer less from dilution, since the egr system is fed through the intake system while the exhaust restriction happens in the cylinder itself.
Obviously, this is over simplified, the work of compression and expansion have to be integrated, there is no consideration of the heat of compression or the differences in heat between the exhaust and intake temperatures. It also assumes that the restriction in the exhaust causes a significant amount to be left in the cylinder when compared to having no restriction. But it does provide a viable theory that the intake contamination could cause a decrease of the heat loss from the system, which would make it more efficient, at least in theory.
I think the reality though is that the gains shown on the tc cars is more a function of cam timing and other modifications